Re: [gentoo-user] Lightweight httpd
Eamon Caddigan [EMAIL PROTECTED] writes: [ ... ] Some programs that look interesting (found 'em all in portage) are: boa, monkeyd, cherokee, and fnord. Would anybody with experience in one or more of these share their opinion? Are there any others I'm missing? for the record, on an embedded system I have I run boa: ~ # ps aux | grep boa 229 root392 S /usr/sbin/boa 392K is pretty small, right? this is, however, boa compiled with uClibc. -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] OT: Keyboard layouts
Tommi Pirinen [EMAIL PROTECTED] writes: Ciaran McCreesh wrote: If you think 2.6.x kernels are stable, try using an unpatched 2.6.1 kernel with a GB or Japanese (or possibly German) keyboard. Actually, almost any non-english keyboard is affected somehow. in X at least, my no-latin setup works like a charm. in console, well, I get surprised every time the layout _isn't_ US. =) -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] IDE RAID 0 and 1
Collins Richey [EMAIL PROTECTED] writes: On Tue, 13 Jan 2004 20:20:40 +0100 Jimmy Rosen [EMAIL PROTECTED] wrote: Kernel SW raid: One customer wanted a very cheap storage solution for the cluster I built him. So I deployed a 2x120GB raid 1 on kernel SW raid. Very simple setup, very simple management, good general cheap solution. How was performance in comparison to the Hardware alternatives? for RAID1, kernel software RAID is just as good as hardware. modern CPUs are a _lot_ more powerful than what you find on the cheap RAID1-cards, and the overhead isn't noticeable. another upside, with software RAID1 you can unplug any of the drives, install it on another system without having to think about RAID controllers and mount the drive. that last part is the reason we chose software RAID1 for the servers that have it here. RAID0 and RAID5 are different beasts though. -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] naming an xterm
Andrew Gaffney [EMAIL PROTECTED] writes: I routinely have atleast 3 xterm's at any one time. I need a way to easily distinguish between them. Is there a command I can run from the xterm to change its title to something I specify? personally, I like having a very clean prompt and keep the relevant information about the shell in the titlebar, hence: case $TERM in linux) declare PS1='\w \$ ';; *) declare PS1='\[\e]0;[EMAIL PROTECTED] (\t \d) \w\a\]\$ ';; esac this gives me titles like: [EMAIL PROTECTED] (hh:mm:ss day month day-of-month) path and a prompt consisting of $ . =) -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Kmplayer
Ernie Schroder [EMAIL PROTECTED] writes: [ ... ] It seems that now I have gmplayer which works properly, though I have no idea where it came from. gmplayer is the standard GUI for mplayer. it comes from mplayer being built with --enable-gui. -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Filesystem couldn't be fixed?
(sorry for the late reply, I haven't had time to read my gentoo-user-box as of late) Spider [EMAIL PROTECTED] writes: Terje Kvernes [EMAIL PROTECTED] wrote: giving /usr a partition of its own isn't that odd, is it? :-) Not really, I find it a bit hard to predict how large it shall be though. this is why you have resize_yourfavoritefs and LVM. then again, my home server currently has around 20 partitions. of these, seven are system-related: /, /boot, /opt, /usr, /local, /var, /var/tmp. the rest of the partitions are data-related, be it music, cvs, tftproot, chroot-environments, images etc. and yes, I do of course run LVM with all of this stuff. Hmm, okeis, Makes me wonder, how much space do you waste on such a setup to make it accept the constant bloating of software? I don't waste much space at all, since I keep the filesystems around 85% full or so and extend on demand. and with todays harddrives, even if it costs me a couple of GiB here and there, it isn't an issue at all. -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] etc-update versus Manual update opinions..
brett holcomb [EMAIL PROTECTED] writes: If you blindly say update it then etc-update sure will mess it up. on a unix-system mostly anything you do blindly, as root, will mess up your system. and, on a unix-system, with your eyes open, that mess can be fixed. Every update that etc-update has wanted to do has wanted to replace my /dev/... with /dev/BOOT and take out my stuff and that would sure hose the system. Other than running it through an editor manually I don't know of anyway to let etc-update do it. you select the option of keeping the original. in the rare event that I actually want a merger, I look at the original, merge the new file, and copy over my local changes. apart from ssh-setup, conf.d-files and make.conf, this happens _very_ rarely. I suppose you could try an interactive update but for files like fstab I'll do it by hadn. the interactive update is pretty decent actually, but I don't use it much myself. -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Filesystem couldn't be fixed?
Spider [EMAIL PROTECTED] writes: [ ... ] Thats strange, 40Mb for root? Unless you partitioned /usr off I'm really quite curious about how you managed to do that. giving /usr a partition of its own isn't that odd, is it? :-) then again, my home server currently has around 20 partitions. of these, seven are system-related: /, /boot, /opt, /usr, /local, /var, /var/tmp. the rest of the partitions are data-related, be it music, cvs, tftproot, chroot-environments, images etc. and yes, I do of course run LVM with all of this stuff. -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] auto Emerge world
Stroller [EMAIL PROTECTED] writes: On Saturday, 23 August 2003, at 11:42 pm, Terje Kvernes wrote: I've been running ~x86 on three boxes for a couple of months now, with the latest and greatest. things work surprisingly well, with some tidbits here and there requiring testes... I think you're saying it takes balls to run ~86..? nah. having a bit of unix know-how helps greatly, but it has been a very smooth ride. recently some issues have appeared with permissions to /dev/null and distcc reacting badly while working with userpriv, but that's about it. -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] auto Emerge world
Collins Richey [EMAIL PROTECTED] writes: [ ... ] (0) By all means use mostly stable packages, but let all update recommendations age 2 weeks before performing the updates, then do them singly and manually watching for anomalies and recommendations. I've been running ~x86 on three boxes for a couple of months now, with the latest and greatest. things work surprisingly well, with some tidbits here and there requiring testes. nothing critical has gone wrong though. The gentoo quality assurance process is good but far from perfect. More than occasionally packages get marked as stable and then must be masked when early adapters discover the rest of the story. shit happens. :-) -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Upgrading to 1.4?
Spider [EMAIL PROTECTED] writes: [ about upgrading ] From pre 1.2? Could be messier but should appear the same way as 1.2 to 1.4 I'm currently doing a 1.0 - 1.4 upgrade on one box. some trickyness, since I need to keep it usable (ie: X has to work, as well as fvwm2, galeon, and some other stuff). it's doable, but the 'emerge -e world' would make things icky, since stuff temporarily would break. the basic idea was to upgrade in steps, making nice graphs in your head with dependencies. the worst one to catch was gtkmm, which required a rebuild of all its dependencies in a certain order -- which didn't get caught by Portage. then again, I wouldn't expect it to, since things were installed and working. :-) -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] bttv where is it?
(subject lowercased) Meka[ni] [EMAIL PROTECTED] writes: I've seen that my friend has no problem with bttv, but I can not find it. I've compiled everything under v4l but there is no bttv module. My kernel is 2.6.0-test3. Where to find that module? it's part of the kernel, under Multimedia devices - Video for Linux Video for Linux - BT848 or so. -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] gentoo-sources
Chris I [EMAIL PROTECTED] writes: [ ... ] I had to go through a bit of trouble to get my synaptics touchpad working. It wasnt hard, it was just different than with 2.4 -- it threw it into a ps/2 compat mode, which i thought was how the device normally worked, wheras I had to change it to use a different mouse driver and protocol. bah. lucky bugger, I'm one of the people who have the issue of Synaptics support being detected, but not being there. I had to remove Synaptics support from the kernel to make things work. :-/ and yes, I will eventually try to produce some constructive testing for lkml, it's just that the laptop is currently my main box, and I hate booting it... -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Best processor for Video Encoding type job
[EMAIL PROTECTED] (Andrew Farmer) writes: [ ... ] Anyway, if you're planning to do serious video work, you probably ought to use OS X anyway. (Horrors!) As great as Linux is, I don't think there's a lot of movie editing software for Linux, other than perhaps film-gimp; however, OS X has a lot of media software available. have you tried Kino? some of my users are very happy with it, but it probably depends on your needs. -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Problem with blocked ebuilds
Ben Ricker [EMAIL PROTECTED] writes: I forgot to mention that I had tried that and perl will not even update. It fails with the same error. unmerge the blocking packages, then update perl. -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] PHP or Perl
daniel [EMAIL PROTECTED] writes: [ ... ] for example, mail() is a php function that sends mail to someone. this capability is available in perl, but requires much more work on your behalf. in perl, you have to open a pipe to your mail binary, write to it and then close the pipe. eh, no, if you're sending a mail with Perl, you'll probably want to do: perl -MCPAN -e 'install Mail::Mailer' and read the documentation for that module. i love perl. it's the shiz-nit when it comes to writing log-parsers, or system administration tools, but for my money, php is just faster and easier to use for the web. with the CGI module, good abstractions and a good plan, perl is extremely nice for the web. the main problem with people writing web applications is that they seem to think that printing out HTML with print is something you'll need to do here and there in every other script. to deal with our machine database at work, we have something that looks a lot like this: my $page = new HTML::Page; $page-attribute('title', 'My title'); $page-attribute('header', 'First header (automatically h1)'); $page-attribute('leadin', 'The first teaser paragraph'); my $part = new HTML::Page::Part; for my $node (@hostlist) { next unless $node-attribute('retired'); # reblessing is also an option. push @nodelist, new HTML::Host $node; } @nodelist = sort { $a-attribute('retired') = $b-attribute('retired) } @nodelist; $part-attribute('header', 'retired hosts (automatically as a h2)'); $part-attribute('nodelist', @nodelist); $part-attribute('keylist', qw|name jack retired|); $page-add_part($part); $page-display() you construct a page as a series of parts, where each part has a given set of characteristics and some data in it. the Part and Page modules both inherit a bit of HTML-crud so they can produce valid HTML or XHTML depending on the use. for our tasks, CSS does all the formatting. but, the point is more that the CGI (which isn't formally CGI due to mod_perl) doesn't contain a single piece of HTML. achieving this with PHP and include is butt ugly, but it will get less ugly when PHP finally gets a somewhat decent object system. [ ... ] -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] upgrading the kernel
Ciaran McCreesh [EMAIL PROTECTED] writes: [ ... ] Again, no. All you should have to do: * emerge sys-kernel/development-sources * emerge module-init-tools (from ~x86) [ ... ] -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] SUMMARY: Kernel 2.6.0 test1 Exploits
Tom Syroid [EMAIL PROTECTED] writes: [ ... ] - anyone know if I have to manually create the /sys directory? Bug? [creating the /sys directory removes the error -- tested as I was writing this] you need to create /sys as it is the mount point for the filesystem, just like you'd create /dev and /proc. - apparently the init scripts are looking for the ../kernel-2.4 module dir. Bug? probably. I haven't looked at it. - I compiled SBLive support direct into the kernel -- no joy. No sound. compile modular sound support, build ALSA modules, install ALSA from portage. it works as a charm here. - the current nvidia kernel will not compile against 2.6.0 -- you'll have to use the default XFree nv driver. this has been solved, of sorts. although I've seen complaints about stability. it should work with module-init-tools, see below. - If you've got your mouse threshold up (mine was at 6), cursor movement under 2.6 is erratic and skitzy. Drop it back to the default of about 3 and you'll be good to go. heh. as long as you don't have a Synaptics-broken touchpad, and actually have a mouse, yeah. :-p All else appears normal at this time. More as I learn more. with regards to modules, I see you got prodded to module-init-tools from ~x86. why this is in ~ is a bit besides me. [x200 /] # grep KEYWORDS /usr/portage/sys-apps/module-init-tools/module-init-tools-0.9.12-r1.ebuild KEYWORDS=~x86 ~amd64 ~ppc ~sparc ~alpha ~mips ~arm -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] mplayer stalling after a second
Henti Smith [EMAIL PROTECTED] writes: Terje Kvernes [EMAIL PROTECTED] wrote: Henti Smith [EMAIL PROTECTED] writes: I have a problem with maplayer lately. It used to work fine, but now when I try to play a file it playes for a second then stops. I can forward and reverse fine and it playes for a second then stops again what output plugin are you using? I've tried them all .. same result does strace -f mplayer testmovie.mpg show anything interesting as mplayer stops? -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Re: Windows partitions ro
Jorge Almeida [EMAIL PROTECTED] writes: [ ... ] /dev/hda8 /mnt/winvfatrw,user,umask=000 0 0 my personal solution is to use autofs and generate auto.misc for /misc with the proper permissions. the snippet below is from a RedHat box at work, but I do the same under Gentoo at home. $ grep card /etc/auto.misc lcard -fstype=auto,user,nosuid :/dev/sda1 card -fstype=vfat,user,nosuid,quiet,uid=3962,gid=13962,umask=077 :/dev/sda1 this is generated from around 100 lines of pretty perl and called from GiveConsole. -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Mad at Mozilla 1.4 : it broke my galeon.
Jonathan C. [EMAIL PROTECTED] writes: [ ... ] Does anyone have any suggestions as to how make my dear galeon work again? edit /usr/portage/profiles/package.mask, and remove the line about galeon being masked, after that: USE=gtk2 emerge mozilla galeon that works fine for me. -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Opteron anyone?
Jerry McBride [EMAIL PROTECTED] writes: [ ... ] A small thing ti consider also, the fastest production opteron is only a 1.8gig chip. Wait till the xeon stompers come out this fall... clock frequency doesn't equal the ability to crunch numbers. on a RISC, this would be a fair assumption. CISC has by far tossed this idea out the window. just look at standard Athlons compared to P4, or even look at what an Alpha at 400Mhz can do. -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Kernel 2.6 Test
(reference header fixed manually. when starting a new thread, please don't reply to an existing article.) Tom Syroid [EMAIL PROTECTED] writes: I'm preparing to build/test the new 2.6 kernel on my dual Athlon box. I'm current running gentoo-sources 2.4.20-r5. welcome to the club. My question is this: In the past it's been noted on numerous occasions the importance of having the link /usr/src/linux pointing to the current kernel before configuration. Does this advice also apply to the 2.5.x series kernels, linked as linux-beta? generally speaking, you want /usr/src/linux to point to your running kernel. some ebuilds use /usr/src/linux and not /lib/modules/`uname -r`. In other words, do I need to create a /usr/src/linux symlink to the 2.6 test directory, or will the existing linux-beta symlink suffice? I used /usr/src/linux even for my 2.5.X-testing because of ebuilds doing interesting things at times. on the upside, things work. :-) -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] mplayer stalling after a second
Henti Smith [EMAIL PROTECTED] writes: I have a problem with maplayer lately. It used to work fine, but now when I try to play a file it playes for a second then stops. I can forward and reverse fine and it playes for a second then stops again what output plugin are you using? -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] [OFF-TOPIC] How can i test my Bandwidth ?
Ricardo Nuno [EMAIL PROTECTED] writes: can you be more specific? hm, I though we'd have an ebuild for it, sorry. what is TTCP and here can i get it. TTCP (Test TCP) has become more a concept than a single application these days. the basic idea is that you put up a sender and a receiver and see how much data your connection can move. it is only usable if you have access to machines on both ends of the connection you're looking to test. keep in mind that you'll want to have few hops between your boxes, you don't want your ISPs backbone to another country to affect the results. one of the original versions can be found here: url: ftp://ftp.sunet.se/pub/network/monitoring/ttcp/ . there are also java versions out there in the wild. -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] HELP! I dorked it myself!
Spider [EMAIL PROTECTED] writes: [ ... ] boot with init=/bin/sash very sound advice, and thank you Gentoo for having /bin/sash installed by default. :-) mount -o remount,rw / you probably want to run 'devfsd' at this stage. mount /usr/ emerge baselayout if mount fails, and you're unable to run devfsd, follow the advice Bryan D. Stine gave in using full devfs names of partitions when mounting. -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] 2.6.0-test1 badness
Kurt V. Hindenburg [EMAIL PROTECTED] writes: Here's something scary... Linux rachael.cn.net 2.6.0-test1 #2 Tue Jul 15 11:58:53 EST 2003 i686 AMD Duron(tm) processor AuthenticAMD GNU/Linux Jul 15 12:50:57 [kernel] Badness in pci_find_subsys at drivers/pci/search.c:132 Jul 15 12:50:58 [kernel] Badness in pci_find_subsys at drivers/pci/search.c:132 Jul 15 12:51:04 [kernel] Badness in pci_find_subsys at drivers/pci/search.c:132 Jul 15 12:51:05 [kernel] Badness in pci_find_subsys at drivers/pci/search.c:132 Jul 15 12:51:11 [kernel] Badness in pci_find_subsys at drivers/pci/search.c:132 Jul 15 12:51:12 [kernel] Badness in pci_find_subsys at drivers/pci/search.c:132 Jul 15 12:51:18 [kernel] Badness in pci_find_subsys at drivers/pci/search.c:132 Jul 15 12:51:19 [kernel] Badness in pci_find_subsys at drivers/pci/search.c:132 this has been reported on linux-kernel earlier. can you test to see if this also happens with 2.5.74 or 2.5.75? -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Update world, new problems with modules
Michael W. Holdeman [EMAIL PROTECTED] writes: Did anyone solve this yet? I messed up BIG time and unmerged baselayout on my desktop! this shouldn't be such a big deal? boot from the installation CD, mount your installation under /mnt/gentoo as usual, and chroot to /mnt/gentoo. after that, emerge baselayout. etc-update might be boring afterwards, but take care and you should be fine. [ ... ] -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Kernel 2.6 net drivers
Andrew Gaffney [EMAIL PROTECTED] writes: Is there no longer support for Davicom NICs using the dmfe driver? I can't seem to find it anywhere in menuconfig. I even searched through the .config and can't seem to find anything in the net section that has anything to do with it. since Davicom chips are a crappy tulip clone, support has been moved to the tulip driver. you want to check 'Tulip family network device support' and pick the appropriate driver from the selection you then get. for a linux-kernel reference, look here: url: http://groups.google.com/groups?selm=linux.kernel.20020519105614.A10528%40gtf.org (google is a good friend) -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Opteron anyone?
Jerry McBride [EMAIL PROTECTED] writes: [ ... ] Dan, can you do me a favor? Next time you boot the opteron, make a note what the bogomip number is and let us know... I'm curious what it reports. bogomips aren't worth anything in real life. they tend to reflect the clock frequency of the CPU and little else. at work we've just ordered a dual 1.8Ghz Opteron to do numbercrunching. we've also bought a dual 3.06Ghz P4 Xeon with HyperThreading to compare it with, but there is very little doubt that the Opteron will come out first. have a peek at this review from Toms Hardware: url: http://www6.tomshardware.com/cpu/20030422/index.html (sigh, I three more days until the boxes come to me. but they won't be running Gentoo though.) [ ... ] -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] lots of duplicates in /usr/lib
Thomas Schweikle [EMAIL PROTECTED] writes: Hi! on one system I have found lots of duplicate libs in /usr/lib: -rw-r--r--1 root root 1303154 May 25 02:48 cracklib_dict.pwd -rw-r--r--1 root root 1303154 May 25 02:48 cracklib_dict.pwd ... -rw-r--r--2 root root 561 Apr 28 22:42 gnomemmConf.sh -rw-r--r--2 root root 561 Apr 28 22:42 gnomemmConf.sh ... -rw-r--r--2 root root 1006 Mar 26 19:22 libBrokenLocale.a -rw-r--r--2 root root 1006 Mar 26 19:22 libBrokenLocale.a ... -rw-r--r--2 root root37294 Jan 27 16:14 libORBitCosNaming.a -rw-r--r--2 root root37294 Jan 27 16:14 libORBitCosNaming.a ... there are 307 duplicates found within /usr/lib (using ls /usr/lib ! uniq -d ! wc -l). Any idea where these come from? file system corruption? what does 'ls -i' tell you about the identical files? -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Opteron anyone?
Daniel Robbins [EMAIL PROTECTED] writes: On Mon, Jul 14, 2003 at 03:22:06PM +0200, Terje Kvernes wrote: ordered a dual 1.8Ghz Opteron to do numbercrunching. we've also bought a dual 3.06Ghz P4 Xeon with HyperThreading to compare it with, but there is very little doubt that the Opteron will come out first. I wouldn't make that assumption. I could very easily see the dual Xeon outperforming the dual Opteron -- make sure to test the specific application first. In fact, I would wager that the dual Xeon will outperform the Opteron in most situations, based on my experience. we're mostly talking about locally produced pieces of Fortran running over rather large datasets for a few days or weeks at a time. we're looking at mostly floating point and a few other tidbits. from what I've seen of similar tests, there is little doubt that the Opteron will be a nice tool for the task. The Opteron 240 seems to be performing somewhere between an XP 2000+ and XP 2200+; not even up to Athlon MP levels yet -- based on my testing, which is by no means exhaustive. And hyperthreading can make a very significant difference for many types of loads, too. for our general case, HyperThreading gives almost nothing in the best case. worst case, it actually costs us some percent here and there. this is, of course, a known issue. we're not talking about desktops running your-favorite-WM here. :-) Be sure to benchmark what you are actually going to be using the machine for. A lot of us work under the assumption that 64-bit is faster, and that's not necessarily true. that much we know. but, there is a reason why you'll still see people running certain calculations on old Alphas, because they're still the best ting for certain things. Opteron isn't a desktop winner. it's a decent performer for such tasks, but not a winner. for certain types of numbercrunching though, all tests I've seen point towards a nice big boost. Portland Group has some small tidbits about 64bit Fortran if one is interested. :-) (and no, I don't do Fortran, I just like to test hardware. :-) -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] gentoo is slow
Christian Aust [EMAIL PROTECTED] writes: [ ... ] p.s.: I'd like to know because I'm just about to switch one of our production machines over to gentoo... so server performance matters to me. apart from the standard warnings around I hope you know what you're doing when you're putting Gentoo onto production machines, you do *not* want a preemptive kernel on a heavy server. -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] gentoo isn't slow :)
Daniel Robbins [EMAIL PROTECTED] writes: [ ... ] I think that a modern uni-processor system with preempt and the low-latency patches enabled would be great for audio work. I don't think you'd specifically need SMP, unless maybe I'm unaware of some special situations where it would be needed. I'd agree with this, but since I mostly live on dual systems I'd just like to say that you _know_ you're on a dual box when you do a lot of stuff. :-) it's very nice to be able to emerge world, watch a movie, surf with galeon, read news over ssh and never ever feel latency or skipping anywhere -- on a dual 450Mhz. the worst culprit is actually flash, which I need to have installed for a few really annoying reasons. a single 600Mhz I have doesn't yield the same carefree environment, even though it runs 2.5.7* with all the trimmings. if you're just doing audio, a single CPU will be fine. if you're planning on a couple of other CPU-intensive things at the same time, well, look at your wallet and decide. :-) -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] 2.5.74 kernel with alsa
Collins Richey [EMAIL PROTECTED] writes: Sven Blumenstein [EMAIL PROTECTED] wrote: Enable OSS emulation in your kernel... thats what I did with 2.5.70 and it works fine :) Yeah, the old should be ok. I have all the necessary modules and settings and modules-update, but no joy. [ ... ] snd_ens137113572 0 snd_rawmidi14720 1 snd_ens1371 snd_seq_device 4100 1 snd_rawmidi snd_pcm62592 1 snd_ens1371 snd_page_alloc 4612 1 snd_pcm snd_timer 15872 1 snd_pcm snd_ac97_codec 39428 1 snd_ens1371 snd30724 7 snd_ens1371,snd_rawmidi,snd_seq_device,snd_pcm,snd_timer,snd_ac97_codec Any ideas? modprobe snd-pcm-oss? -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] kerenl 2.5.74 fails to boot
Collins Richey [EMAIL PROTECTED] writes: [ ... ] No framebuffer support, but the ^%$£ 2.5 config turned off VGA console support! I've had the same problem over and over again with 2.5.*, with or without VGA console support. of course, 2.5.74 fails upon building spb2 (FireWire storage support), so I'm back to 2.5.70. a small laptop is a funny test candidate for unstable kernels. :-) [ ... ] -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] File Sharing between Linuxes
Chris I [EMAIL PROTECTED] writes: [ ... ] The only problem is that there seems to be a bug when mounting an nfs share on localhost that I cannot fix. It takes five minutes, yet works fine after mounting. have you tunneled the appropriate portmap stuff as well? -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] convert unix time stamp to human readable
Timo Boettcher [EMAIL PROTECTED] writes: [ ... ] == Begin paste unix-to-human-time.pl == #! /usr/bin/perl -p s/^\d+\.\d+/localtime $/e; == End paste unix-to-human-time.pl interesting use of $. :-) -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] OT: Windows to linux... What was that app?
Zack Gilburd [EMAIL PROTECTED] writes: [ ... ] Amen. Sysadmins are the downfall of both operating systems. Unfortunately, it's a lot easier to get behind in patches (unless you are using some nifty package management like portage ;)) in a Linux distribution than it is in Windows 2003, for example. almost every modern distro has an online update tool. I'd like to actually hear about one that _doesn't_ have this. with that being said, with regards to parents, family or friends, most of the people I know come in two categories. they either know what they're doing and run linux just fine, or they don't know what they're doing and need a lot of help anyway. in which case I can help people with linux, but not with Windows. hence, people around me generally run linux. my father does however run Windows -- mostly because he's a manager who deals with PowerPoint, Excel and Word. last time I worked with him though, his staff ran Linux out of choice and he was fine with that. he also quotes well, sends plain text emails and makes it rather painless to communicate with him. all servers and such things around him tend to run linux though. :-) my father also occasionally complains about Windows Update and bug fixes or service patches that breaks things. when this happens with Gentoo with people I know ssh is usually all the cure we need. at ork[0] we only have linux-boxen, and I must admit it's a nice feeling when a 60-something year old woman sends a mail and wonders why Evolution won't sync her palmpilot anymore. :-) (the answer is usually that pam.d and console.perms aren't friends anymore. we run RedHat at ork[0].) [ ... ] -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Windows to linux... What was that app?
Dan Fairs [EMAIL PROTECTED] writes: [ ... ] Public sector adoption in particular is interesting. Here in the UK, it looks like the government are a Microsoft shop, with the Government Gateway and all. Has anyone been involved in a Gentoo deployment in the public sector? at the department of Mathematics at the University of Oslo we use Linux for both servers and clients, but we have access to a single Windows terminal server via rdesktop. the linux distribution of choice around here is RedHat. we also have some OS9 macs, but we're migrating over to OSX slowly. the rest of the university has a rather large amount of linux boxen. I can't give an exact number, but other departments (apart from the obvious Computer Science) are also major linux users. -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Server application choices?
Matthew Daubenspeck [EMAIL PROTECTED] writes: [ ... ] Although, I just use ssh to copy files (sftp or scp). 'rsync -e ssh' is your friend. :-) -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Problems compiling modules
Jesper Blauendahl [EMAIL PROTECTED] writes: [ ... ] Failed to calculate dependencies modprobe: QM_MODULES: Function not implemented What am I missing in the setup ? loadable module support? -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Problems compiling modules
Martin Schlemmer [EMAIL PROTECTED] writes: On Mon, 2003-06-30 at 09:08, Jesper Blauendahl wrote: I'm trying to switch to the development-sources (2.5.70) to get a better support for my hardware. The kernel compiles nicely. But I have SCSI Emulation, SCSI device support, SCSI CD-ROM support, and SCSI generic support set to compile as modules, and when I execute 2.5.70 is fairly old. Rather try 2.5.73 (or -bk6), as a few things have been fixed. on my X200 2.5.70 is the only kernel I get to actually boot. 2.5.7{1,2} as well as 2.5.66 and 2.5.55 all fail to boot. as soon as I find out what causes this, I'll report it. :-/ -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] how to use a worldfile copied from other system.
Sigurd Stordal [EMAIL PROTECTED] writes: I had to reinstall my gentoo, but I saved my world file, no I would like to use the old worldfile and install all the packages I had installed before. But when I copied this over the worldfile in /var/cache/edb and run emerge -p it complain about the files not installed not being installed. Is there maybe another place I should put the worldfile, or how do I get it to install as new all the files in worldfile that are not previously installed. emerge -p $( old_world_file ) -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] IDE RAID suggestions
Matt Tucker [EMAIL PROTECTED] writes: [ ... ] If you've got the money (and particularly if you want to do RAID-5), I highly recommend the 3ware cards. They've got a 4-channel card (so you can put each drive on a separate channel, with none of this master/slave stuff) which supports RAID-5 (as well as 0, 1, etc. of course), has open-sourced drivers (included in the kernel) and acts as a single scsi interface (per array). at work we've recently tested the 8-way 7500-series controller from 3Ware. so far it just works. okay, so since Serial ATA (that would be the 8500-series) drives are still in the blue we have 9 IDE-cables in the cabinet. a bit of a trick to get it tidy, but still. performance is good, no problems at all so far. /dev/sda1 reiserfs1.3T 373G 957G 28% /mnt/testraid it's also fun to see T in the output from 'df -h'. :-) -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Booting Gentoo from READONLY medium
Michael Maeder [EMAIL PROTECTED] writes: [ ... ] So I shrinked Gentoo to fit on a 64MB CF card (can be shrinked even a lot more). this is interesting, very interesting. can you give some hints as to how you did this? I tried a bit, but gave up. the biggest issue I had was that I _really_ want to use uclibc instead of glibc. [ ... ] -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Ext3 LVM
Terje Kvernes [EMAIL PROTECTED] writes: Matthew Kennedy [EMAIL PROTECTED] writes: [ ... ] Well at first it seems like a solution, but I don't think it is. it should work out-of-the-box on ext3 as well as ext2, don't let the name fool you. :-) gaaah. my sincere apologies. the frigging filesystem I tested this on was mounted as ext2 even though it was formatted ext3. solutions are, again, out in the blue. [ ... ] -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Ext3 LVM
Matthew Kennedy [EMAIL PROTECTED] writes: [ ... ] XFS has xfs_growfs which works on mounted volumes. I *think* resierfs has something similar. ext2 has it as well. Reiserfs has 'resize_reiserfs', which works well. by the way, shrinking Reiserfs is not something you'd really want to do -- as it rebuilds the internal tree, which takes _time_. I thought of shrinking a 450GiB volume by 50GiB or so. estimated time was around 26 hours or so. :-) -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] LSB compliance
Daniel Carrera [EMAIL PROTECTED] writes: converting a binary rpm to an ebuild is doable. but only to the extent of minor dependency-checking and tracking installed files. maybe someone feels like a masochist and fixes this sometime. There's no need to fix anything. As long as an RPM can get installed, the LSB doesn't care. I care. :-) I'd like to be able to install / uninstall the rpm via emerge and have emerge track its files. an alien for emerge if you like. in short, the FHS and LSB do not produce a solution for the problem. it's more defined as a non-issue, and you get pointed to /opt -- which according to the FHS doesn't fit either Gnome or KDE anyway. Why doesn't /opt work? I've looked at the FHS section on /opt and I don't see what's wrong with /opt/gnome and /opt/kde. because stuff in /opt is supposed to be static packages, not dependent on anything else on the system, least of all other packages (heck, formally even depending on glibc is bad). this also works the other way around, no packages should ideally depend on anything located in /opt. both Gnome and KDE break this both ways. they depend on stuff in / or /usr, and other packages depend on KDE or Gnome libraries. -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Ext3 LVM
Carlos Molina Garcia [EMAIL PROTECTED] writes: [ ... ] At this moment, I can't find any utilitie for ext3 that permit resize the lv on a mounted fs. On the reiser area, the reisize utilitie can't be used on a mounted fs. eh? resize_reiserfs works fine on mounted volumes. -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] LSB compliance
Ted Goodridge, Jr [EMAIL PROTECTED] writes: /opt is bad because it's contents are not supposed to be dependent on anything else. /usr/local is bad because it's meant for non-distribution packages. I'm curious...why is it so much of a problem to put it in /usr/local? url: http://www.pathname.com/fhs/2.2/fhs-4.9.html I am not that familiar with FHS or LSB, so please forgive me if this sounds totally nuts. well, then read the FHS. :-) On my openbsd machine, they have decided /usr/local is for anything that is not essential to the base system. To their way of thinking, anything besides the base system (perl, libs, ssh, etc) is considered a non-distribution package. that's OpenBSD. not Linux. personally I find it totally broken that they separate out their ports tree as something other than system software. this is why we have /usr. this is of course also annoying since your basesystem is _not_ maintained via ports. how then do you do the equivalent of emerge --update glibc? anyhow, the issue is more the definition of local software. OpenBSD uses the term as stuff installed apart from the core stuff we deem that you have to have. Linux defines local as whatever the sysadmin installs apart from what the distribution provides. for Gentoo this is even more true than ever. the question then comes of course, where do you install your local software on OpenBSD? /local? which by the way is not defined in the hier (7) manpage of OpenBSD. and if you use /usr/local, what happens if you install a port that has files that collide with what you've installed on your own? Why do KDE or GNOME _not_ fall into this category? that's nowhere near the point. KDE and Gnome are installed by a packager coming with the distribution. that makes /usr/local a big no-no. -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Opinions on Gentoo
Daniel Carrera [EMAIL PROTECTED] writes: [ ... ] Regardless, I'm glad to know that ebuilds are easier to make than RPMs. Who knows, I might actually need to make one some day. just have a peak at how ebuilds work and look: url: http://www.gentoo.org/doc/en/gentoo-howto.xml url: http://cvs.gentoo.org/cgi-bin/viewcvs.cgi/gentoo-x86/ as an easy example, look at the latest sash build: url: http://cvs.gentoo.org/cgi-bin/viewcvs.cgi/*checkout*/gentoo -x86/app-shells/sash/sash-3.4-r5.ebuild?rev=1.16 [ ... ] What exactly is the problem with Gentoo's LSB compliance? as others have pointed out, the LSB has a strict idea about what directories you can have under /usr. Gentoo has a directory 'kde' located under /usr, to make it possible to have several different versions of KDE installed at the same time. when LSB tells the developers how to solve this differently, without breakage, I'm sure they'll move and port stuff. but I wouldn't hold my breath for the LSB to fix support for issues like this. -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] XFS not enabled, is that on purpose? JFS users beware of XFS.
Oleg Letsinsky [EMAIL PROTECTED] writes: [ ... ] Strange. I use XFS from the day it came to Linux and have never had a problem with it - even my experiments with early nvidia drivers, causing hardware lockups (during compilations) didn't break it. filesystems are interesting things. I've had XFS, heck even ext3, crash on me, but never had problems with Reiserfs. other people have _very_ different tales. but, in all honestly, who has _statistically_ significant material worth talking about? if XFS fails on a box or two during a year, is it the filesystem, hardware, kernel, userspace or what? how many combinations and trails have people run? consistently? -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] LSB compliance
Daniel Carrera [EMAIL PROTECTED] writes: [ ... ] With respect to (1), I don't think that's quite true. I read an article a few days ago by one of the LSB members who said that the LSB doesn't require distros to use RPM as their package system. Rather, it requires them to be *able* to install RPMs. He gave the example that Debian's alien program fills this requirement. If this is correct, we could simply have an equivalent of alien for Gentoo. converting a binary rpm to an ebuild is doable. but only to the extent of minor dependency-checking and tracking installed files. maybe someone feels like a masochist and fixes this sometime. With respect to (2), I must say I don't really understand the problem. If the Gentoo has KDE in /usr/kde/2 and /usr/kde/3, and LSB doesn't allow /usr/kde, why don't we just move /usr/kde to /usr/share, so we would make: /usr/share/kde/2 /usr/share/kde/3 because that's misuse of /usr/share. from: url: http://www.pathname.com/fhs/2.2/fhs-4.11.html # The /usr/share hierarchy is for all read-only architecture # independent data files. in short, the FHS and LSB do not produce a solution for the problem. it's more defined as a non-issue, and you get pointed to /opt -- which according to the FHS doesn't fit either Gnome or KDE anyway. I'm sure that this question just demonstrates my ignorance of LSB and Gentoo, but heck, asking is the only way I'll learn. I've heard that reading also helps. :-) -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] Via C3 optimization
Patrick Nehls [EMAIL PROTECTED] writes: I use one of the Via mini-itx boards with embedded C3/800Mhz CPU. It runs Gentoo 1.4/current portage tree just fine but as stated, only with the i586 optimizations. I believe the Gentoo user forums had some threads in the past on this. Apparently the C3 does not implement 1 optional instruction in the x686 insruction set and gcc makes use of that in the x686 optimizations. /etc/network/iptables # cat /proc/cpuinfo processor : 0 vendor_id : CentaurHauls cpu family : 6 model : 7 model name : VIA Samuel 2 stepping: 3 cpu MHz : 531.837 cache size : 64 KB fdiv_bug: no hlt_bug : no f00f_bug: no coma_bug: no fpu : yes fpu_exception : yes cpuid level : 1 wp : yes flags : fpu de tsc msr cx8 mtrr pge mmx 3dnow bogomips: 1061.68 quite a few applications are built for i686 when running on this thing. they do seem to work as intended. maybe it's relevant that I'm building and running under a uclibc-environment. this is a Lex CV860A-3R53[1] barebone box that'll be used for a firewall eventually. since I had a 8MiB flashcard around I've installed on that. with busybox and uclibc it's outright scary what one can squeeze into 8MiB. sshd, boa, rrdtool, iptables, strace, pcimodules, rsync and the big prize, perl. okay, we don't get any libraries or packages without manually installing them (the full package-set it 27MiB) but we have the perl binary. heck, I still have a few hundred kilobytes free. :-) url: http://www.lex.com.tw/cv860a.htm -- Terje -- [EMAIL PROTECTED] mailing list
Re: [gentoo-user] portage downgrade?
[EMAIL PROTECTED] writes: Susie [EMAIL PROTECTED]: Is there something wrong with the last portage release? I just did an emerge rsync emerge -up --deep world and got: [ebuildUD] sys-apps/portage-2.0.46-r12 [2.0.47-r2] I'm experiencing the same thing here. Is this a case of one mirror getting ahead of the others? Was 47-r2 unmasked, then remasked again? I recall a similar situation occuring a few weeks ago; I'm awfully curious to know what causes this. url: http://bugs.gentoo.org/show_bug.cgi?id=16047 at least proves that portage 2.0.47 still can break stuff, it might have been pulled back for more testing. which would be a good thing. :-) -- Terje -- [EMAIL PROTECTED] mailing list